Enhancing Forest Inventories Using Lidar and GIS
Enhancing Forest Inventories Using Lidar and GIS Dr. Kevin Lim and Mr. Chad St.Amand P.O. Box 30030 Greenbank North PO Ottawa, ON | K2H 1A3 | Canada
Presentation Outline
•
Background
•
High Level Workflow
•
Value Added Information Products and Tools
•
Cost Benefit Analysis
Background
Workflow
Basic Products
High Level Workflow
Acquire LiDAR Data
Acquire Field Data
Process LiDAR Data Perform Statistical Analysis
Apply Models to Landscape
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Inventory Variables Variable
Abbrev.
Definition
Top Height (m)
TOPHT
Calculated as the average of the largest 100 stems per hectare.
Average Height (m)
AVGHT
Calculated as the average height of all trees
Density (stems/ha)
Density
Number of trees per hectare
Quadratic Mean Diameter (cm)
QMDBH
Basal Area (m2/ha)
BA
DBH2 * 0.00007854
Gross Total Volume (m3/ha)
GTV
Honer et al. (1983) equations
Gross Merchantable Volume (m3/ha)
GMV
Honer et al. (1983) equations
Total Above Ground Biomass (Kg/ha) Diameter Distributions
SUMBIO
Ter-Mikaelian and Korzukhin (1997) equations
DD
Volume, BA, Density by Size Class
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Normalization of Points to Terrain
zveg
∆ zgrd
TILE TIN
∆ = Znorm = Zveg - Zgrd
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Lidar Predictors •
Statistical -
•
Percentiles of height -
•
Mean Standard deviation Deciles (p10 … p90) Maximum height
Canopy density -
d1 … d9 Da: Number of first returns divided by all returns. Db: Number of first and only returns divided by all returns.
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Lidar Predictor Surfaces
Each surface corresponds to a lidar predictor. Cell resolution of 20m (or 400m2 in area).
Apply a mask (optional).
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Statistical Modelling
•
Performed outside of ArcGIS -
SAS R Any other statistical package
Models
Transporting Models 300
250
a) PJ Acutal GMV (m3/ha)
Acutal GMV (m3/ha)
250 200 150 100 50
b) SB
200 150 RMForig
RMForig
RMFmf 100 MF
RMFmf MF 1:1
1:1 50
0
0 0
100
200
300
400
0
50
Predicted GMV (m3/ha) 30
25
c) PJ Acutal GMV (m3/ha)
Acutal Dbhq (m3/ha)
20 15 10 5 0 5
10
15
Predicted Dbhq (m3/ha)
150
200
250
d) SB
25
0
100
Predicted GMV (m3/ha)
20
25
20 RMForig 15
RMForig
RMFmf
RMFmf
10 MF
MF
1:1 5
1:1
0 0
5
10
15
Predicted Dbhq (m3/ha)
20
25
0.35
æ 1 æ Dbh - m ö 2 ö æ æ Dbh - a ö c ö 1 expç - ç expç - ç ÷ ÷ ÷ ÷+ ç è b ø ÷ s 2p ç 2è s ø ÷ø è ø è
Relative GMV
c -1
Sw
0.2
Pj
0.15
Plot
0.1
Mixture
0.05
Weibull
0 10
12
14
16
18
20
22
Diameter class (cm) 0.18
Plot 49
0.16 0.14 Relative GMV
c æ Dbh - a ö f (x) == ç ÷ aè b ø
0.25
0.12
Sb
0.1
Sw
0.08
Pj
0.06
Plot
0.04
Mixture
0.02
Weibull
0 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Diameter class (cm) 0.2
Plot 44
0.18 0.16 Relative GMV
Diameter Distribution Modeling
Plot 20
0.3
Pj
0.14 0.12
Sw
0.1
Bf
0.08
Bw
0.06
Plot
0.04
Mixture
0.02
Weibull
0 10
12
14
16
18
20
22
24
26
Diameter class (cm)
28
30
32
34
36
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Apply Models to Landscape
Value Added Information Products and Tools
Cost Benefit Analysis
A. Inventory Data Acquisition and Treatment Costs Cost comparison items
Cost paid by TRAD
Inventory data acquisition and processing costs
Costs REAL
LIDAR
$/ha
$/ha
$/ha
Lidar data acquisition
Tembec
-
-
1.00
Lidar processing
Tembec
-
-
0.45
Lidar validation plots
Tembec
-
-
0.15
Photo acquisition
OMNR
0.46
0.46
0.46
Photo processing and interpretation
OMNR
0.44
0.44
0.44
Traditional inventory plots
OMNR
0.40
0.40
0.40
$1.30
$1.30
$2.90
Sub-total § §
m2
RMF LIDAR acquisition cost about $0.40/ha for 0.5 pulses/ Cost depends of the size of the acquisition, better price for larger area
B. Forest Operations Cost Analysis • Forest management plan (FMP) revisions • Better wood allocation at the planning stage • Budget forecast • Decrease of Forest and Mill Inventory • Better freshness on the wood products • Feller-buncher productivity (m3/ha) • Full-tree productivity (m3/stem) • Skidding productivity • Wood cutting optimization • Wood damage on immature wood • Wood delivery logistics • Floating costs • Block layout • Handling productivity • Better road location & design • Road construction • Road maintenance • Silviculture funds • Silviculture cost • Indirect costs
Total
N/A $0.13 m3 $0.02 m3 $0.07 m3 m³ $0.09 m3 $0.08 m3 $0.19 m3 $ ??? $ ??? $ ??? $ ??? $0.05 m3 $0.05 m3 $0.02 m3 $0.01 m3 $0.43 m3 $0.03 m3 $0.06 m3 $0.02 m3 $0.15 m3
$1.40 m3
C. Mill Cost Analysis Savings • Sawmill scheduling • Wood purchase • Wood net value - sawmill productivity • Lumber Value
N/A $0.11 m3 $0.12 m3 $0.07 m3
Total $0.30 m3
Scenario Results – REAL vs. LiDAR Cost Analysis 1 Inventory acquisition & processing 2. Forest operations 3. Mill
Total Savings with LiDAR
$1.60 /m3 X 500,000 m3 / year = $ 800,000 / year
Model Cost of LiDAR acquisition and Processing = $1,006,400 Payback = 1.3 years
Kevin Lim | [email protected]
Chad St.Amand | [email protected]
Presentation Outline
•
Background
•
High Level Workflow
•
Value Added Information Products and Tools
•
Cost Benefit Analysis
Background
Workflow
Basic Products
High Level Workflow
Acquire LiDAR Data
Acquire Field Data
Process LiDAR Data Perform Statistical Analysis
Apply Models to Landscape
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Inventory Variables Variable
Abbrev.
Definition
Top Height (m)
TOPHT
Calculated as the average of the largest 100 stems per hectare.
Average Height (m)
AVGHT
Calculated as the average height of all trees
Density (stems/ha)
Density
Number of trees per hectare
Quadratic Mean Diameter (cm)
QMDBH
Basal Area (m2/ha)
BA
DBH2 * 0.00007854
Gross Total Volume (m3/ha)
GTV
Honer et al. (1983) equations
Gross Merchantable Volume (m3/ha)
GMV
Honer et al. (1983) equations
Total Above Ground Biomass (Kg/ha) Diameter Distributions
SUMBIO
Ter-Mikaelian and Korzukhin (1997) equations
DD
Volume, BA, Density by Size Class
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Normalization of Points to Terrain
zveg
∆ zgrd
TILE TIN
∆ = Znorm = Zveg - Zgrd
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Lidar Predictors •
Statistical -
•
Percentiles of height -
•
Mean Standard deviation Deciles (p10 … p90) Maximum height
Canopy density -
d1 … d9 Da: Number of first returns divided by all returns. Db: Number of first and only returns divided by all returns.
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Lidar Predictor Surfaces
Each surface corresponds to a lidar predictor. Cell resolution of 20m (or 400m2 in area).
Apply a mask (optional).
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Statistical Modelling
•
Performed outside of ArcGIS -
SAS R Any other statistical package
Models
Transporting Models 300
250
a) PJ Acutal GMV (m3/ha)
Acutal GMV (m3/ha)
250 200 150 100 50
b) SB
200 150 RMForig
RMForig
RMFmf 100 MF
RMFmf MF 1:1
1:1 50
0
0 0
100
200
300
400
0
50
Predicted GMV (m3/ha) 30
25
c) PJ Acutal GMV (m3/ha)
Acutal Dbhq (m3/ha)
20 15 10 5 0 5
10
15
Predicted Dbhq (m3/ha)
150
200
250
d) SB
25
0
100
Predicted GMV (m3/ha)
20
25
20 RMForig 15
RMForig
RMFmf
RMFmf
10 MF
MF
1:1 5
1:1
0 0
5
10
15
Predicted Dbhq (m3/ha)
20
25
0.35
æ 1 æ Dbh - m ö 2 ö æ æ Dbh - a ö c ö 1 expç - ç expç - ç ÷ ÷ ÷ ÷+ ç è b ø ÷ s 2p ç 2è s ø ÷ø è ø è
Relative GMV
c -1
Sw
0.2
Pj
0.15
Plot
0.1
Mixture
0.05
Weibull
0 10
12
14
16
18
20
22
Diameter class (cm) 0.18
Plot 49
0.16 0.14 Relative GMV
c æ Dbh - a ö f (x) == ç ÷ aè b ø
0.25
0.12
Sb
0.1
Sw
0.08
Pj
0.06
Plot
0.04
Mixture
0.02
Weibull
0 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 Diameter class (cm) 0.2
Plot 44
0.18 0.16 Relative GMV
Diameter Distribution Modeling
Plot 20
0.3
Pj
0.14 0.12
Sw
0.1
Bf
0.08
Bw
0.06
Plot
0.04
Mixture
0.02
Weibull
0 10
12
14
16
18
20
22
24
26
Diameter class (cm)
28
30
32
34
36
Workflow Acquire LIDAR Data
Define Stratification
LIDAR Data
Classify Point Cloud Acquire Field Data For Sample Plots Intersect With Sample Plots
Vegetation Points
Ground Points
Create TIN
Field Data Normalize Points To Terrain
Vegetation Points Per Plot Calculate Forest Variables
Forest Variable Statistics
Perform Statistical Analyses
Regression Models
TIN
Calculate LIDAR Predictors
Normalized Vegetation Points Per Plot
Normalized Vegetation Points
LIDAR Predictors
LIDAR Predictor Surfaces
Calculate LIDAR Predictors
Apply Models to Landscape
Forest Inventory Surfaces
Apply Models to Landscape
Value Added Information Products and Tools
Cost Benefit Analysis
A. Inventory Data Acquisition and Treatment Costs Cost comparison items
Cost paid by TRAD
Inventory data acquisition and processing costs
Costs REAL
LIDAR
$/ha
$/ha
$/ha
Lidar data acquisition
Tembec
-
-
1.00
Lidar processing
Tembec
-
-
0.45
Lidar validation plots
Tembec
-
-
0.15
Photo acquisition
OMNR
0.46
0.46
0.46
Photo processing and interpretation
OMNR
0.44
0.44
0.44
Traditional inventory plots
OMNR
0.40
0.40
0.40
$1.30
$1.30
$2.90
Sub-total § §
m2
RMF LIDAR acquisition cost about $0.40/ha for 0.5 pulses/ Cost depends of the size of the acquisition, better price for larger area
B. Forest Operations Cost Analysis • Forest management plan (FMP) revisions • Better wood allocation at the planning stage • Budget forecast • Decrease of Forest and Mill Inventory • Better freshness on the wood products • Feller-buncher productivity (m3/ha) • Full-tree productivity (m3/stem) • Skidding productivity • Wood cutting optimization • Wood damage on immature wood • Wood delivery logistics • Floating costs • Block layout • Handling productivity • Better road location & design • Road construction • Road maintenance • Silviculture funds • Silviculture cost • Indirect costs
Total
N/A $0.13 m3 $0.02 m3 $0.07 m3 m³ $0.09 m3 $0.08 m3 $0.19 m3 $ ??? $ ??? $ ??? $ ??? $0.05 m3 $0.05 m3 $0.02 m3 $0.01 m3 $0.43 m3 $0.03 m3 $0.06 m3 $0.02 m3 $0.15 m3
$1.40 m3
C. Mill Cost Analysis Savings • Sawmill scheduling • Wood purchase • Wood net value - sawmill productivity • Lumber Value
N/A $0.11 m3 $0.12 m3 $0.07 m3
Total $0.30 m3
Scenario Results – REAL vs. LiDAR Cost Analysis 1 Inventory acquisition & processing 2. Forest operations 3. Mill
Total Savings with LiDAR
$1.60 /m3 X 500,000 m3 / year = $ 800,000 / year
Model Cost of LiDAR acquisition and Processing = $1,006,400 Payback = 1.3 years
Kevin Lim | [email protected]
Chad St.Amand | [email protected]
